Transformer_Magnetic_Core_Issues.pdf

MAGNETIC CORE ISSUES IN POWER TRANSFORMERS AND THEIR DIAGNOSTICS William F. Griesacker and Juan Luis Thierry Doble Engineering Company ABSTRACT General experience in the industry shows that power transformer magnetic cores can provide years of reliable service when properly constructed, and sustain no damage during transportation and installation. However, core problems are experienced in service that require attention. Therefore, recognizing the signs and identifying what action plan is needed to investigate possible problems are critical steps when issues are encountered. Core problems are not usually associated with fast acting failure modes but can deteriorate over time, and their gas generation can mask other possible problem activity in the transformer. In core form transformers, some areas of the core are accessible, while in shell form transformers access to the core is very limited, which makes locating and repairing core problems in the field difficult. The following discussion will present magnetic core problems encountered during manufacturing, transportation and operation of power transformers. This paper will also review the methods used to identify core problems and determine their severity.

INTRODUCTION The transformer core is comprised of many thin sheets of insulated silicon steel stacked together and held in place with a clamping frame in core type transformers. In shell type transformers the core is held in place by the internal frame welded to the tank walls. The core type transformer core is typically comprised of multiple laminations widths that create steps to approximate a circular cross section, which maximizes the quantity of core material inside the circular coils utilized on power transformers. Silicone steel material is selected for its high permeability which provides a low reluctance path to the magnetic flux. Reluctance is analogous to resistance as explained by ohms law, where V=IR and R is resistance. The lower the reluctance, the less opposition there is to flux in the core, which in turn provides an efficient low impedance linkage between the primary and secondary windings. The insulation on the core laminations isolates adjacent steel sheets, minimizing the eddy current path and their losses. The silicon steel and high temperature insulation used on laminations of most modern transformer cores, make the core one of the more durable components of the active assembly that usually requires little to no attention during a transformer’s service life. However, core problems are seen in both the manufacturing setting and in service that can be due to lamination shorting, overheating, core movement or other problems. When core problems do arise, implementing an effective plan to investigate the cause and determine what steps are needed to correct it is an important step to help minimize the down time or the risk of failure of a transformer. A number of tests will be discussed that have shown to be effective in helping identify and characterize core problems; some of these tests were not specifically developed or primarily used for core diagnostics.

CORE MANUFACTURING PROBLEMS Once a transformer has been constructed and successfully tested at the factory, it is presumed to be suitable for service. There is a general expectation that the manufacturer will identify, address and resolve any core problems throughout the manufacturing phase. It also might be presumed that standard factory testing will identify any problems a core may have. In reality, some core defects are difficult to identify and factory acceptance testing, performed in accordance with industry standards, will not catch all

st

© 2014 Doble Engineering Company – 81 International Doble Client Conference All Rights Reserved

of them. Some issues that are encountered in core manufacturing and that are known to cause problems include the following: 

Lamination Edge Damage – Core lamination edges, for discussion purposes, can be broken down into two basic types that relate to the processes used to cut them to their required dimensions. The slit edge refers to the process used to cut laminations to their required width. The sheared, or miter cut edge, refers to the core joint edges and the process used to cut laminations to the required length and end shape. In general, the slitting operation is used to cut a coil of core steel to a manufacturer’s standard width. This coil is then cut, usually using a shearing process, into required lengths for a given core step. Based on factory inspections performed on cores, it appears that the slit edges are more likely to be damaged during manufacturing. This is likely due to the fact that there is much more core surface area comprised of slit edges, and these edges are more exposed and therefore more susceptible to damage. Damage to the core steel edges can create shorts among adjacent laminations, this can cause increased losses in these areas resulting in hot spots on the core. Lamination edge damage can also occur in other areas, for example at the alignment holes. The industry has generally moved away from using core bolts due to shorting conditions encountered in the past. However, holes are commonly used to align the laminations during the core stacking process. If the alignment holes are over stressed during core construction, the lamination edges can be damaged, see examples in Figure 1.

Examples of Edge Damage at Core Alignment Holes Figure 1 Damage to the insulation on the face of laminations can also lead to shorted laminations, especially if damaged areas line up between multiple adjacent laminations in the core and create a path for eddy current circulation between laminations. 

Shorted Laminations – Aside from damage, shorted laminations can occur if conductive debris, hardware, misplaced tools or other conductive items are allowed to come in contact with the uninsulated edges of the core laminations. Improper isolation of the core, primarily from the clamping frame, or insulation that moves can also cause a shorted lamination condition. The more laminations that are involved in a short, generally the worse the problem will be since a larger portion of the flux in the core will be involved in driving current through the shorted laminations. An example of shorted laminations in a shell form transformer is shown in Figure 2.

st

© 2014 Doble Engineering Company – 81 International Doble Client Conference All Rights Reserved 2 of 22

Example of Shorted Core Laminations in a Shell Form Transformer Figure 2 

Excessive Edge Burr – Many power transformer manufacturers purchase rolls of core steel cut to standard widths and have a core cutting operation in the factory to cut the steel to the required lengths and shape to form the core joints. A shear is usually used in this cutting operation. In a well-controlled process, the burr on the shear cut edges is a function of the sharpness and proper adjustment of the shear blade. Regular checking of the burr will give an indication to the machine operator when the blade needs to be serviced. If the edge burr is allowed to become too large it can create shorts between laminations; the losses associated with the core joints will increase and can become excessive.



Improper Core Joint Gaps – Core joints are designed to provide a low loss and low reluctance path for the magnetic flux to pass between adjacent core laminations. The “reluctance path” of a transformer usually refers to the path of a flux loop and its reluctance; for the main flux of the transformer, this path would be confined to the core and the paths which the flux loops travel in it. One critical aspect of the core joint construction is a properly sized gap that minimizes localized saturation of the steel laminations. Core joint gap size is a common check during core construction. Some factors can affect the gap dimension, and problems are sometimes not easily recognized until the top core yoke laminations are inserted. Interference between major components can cause core joint problems, an example is overbuilt coils. In common core form construction, the bottom yoke and limbs of the core are stacked, the core is uprighted and then the coils are lowered on the core limbs. If the coils are overbuilt or grow in height, for example, due to extreme moisture absorption during humid seasons, the coils can interfere with the top yoke assembly. The example in Figure 3 shows a case where the coils grew in height after the sizing process and prevented proper assembly of the top yoke. The manufacturer’s maximum gap tolerance was 3 mm and the gap measured in the left photograph was 6 mm, making the core joint gap well over the acceptable limit.

st


Excessive Core Joint Gaps Figure 3 In shell form transformers, overbuilt winding phase assemblies on some older designs could cause out-of-tolerance gaps in the core. The larger gaps would allow the oil to stagnate in an area with higher losses and temperatures due to the oversized gaps. The oil would decompose and carbonize on the core laminations, as shown in Figures 4 and 5. The left photo in Figure 5 shows localized heating on the core lamination edges between phases as found on this three phase shell form transformer. The right photo shows the effect of core overheating at the miter joints after the external core laminations were removed.

Carbonized Oil Found on a Core Joint Due to Excessive Gaps Figure 4

st


Overheating on Core Joints Due to Excessive Gaps Figure 5 

Out-of-Tolerance Dimensions – Power transformer cores are typically built by laying a lamination flat and stacking laminations on top, as the laminations are laid in place the thickness of the core increases. This thickness of the core is called the “stack height” dimension. The core of a typical power transformer can be constructed from thousands of individual laminations. If the thickness of laminations is out of tolerance, the final core stack height dimension can subsequently be out of tolerance. Small variations in thickness of the lamination steel, the insulation, waviness of the laminations or other causes can result in less core steel placed in the core. If less steel is in the core than the design calls for, the flux density of the core will be higher which can lead to operating closer to saturation and increased core losses. The cut dimensions of core laminations must be precise so that gaps of the miter joints are minimal and uniform across the span of each joint. As laminations are stacked on the core, alignment is critical to meet the design dimensions; any items out of square or shifted can result in unwanted gaps at the joints. Both out-of-tolerance lamination dimensions and out-of-square conditions during building of the core can cause excessive joint gaps or joints that are too close together, which can cause a buckling type wave shape in the lamination steel. An example of a buckling type wave in core top yoke laminations is shown in Figure 6. These problem conditions are often noticed once the top yoke is being installed since this is the final item to complete the core magnetic circuit of a core form transformer.

Example of Wave in Core Yoke Laminations Figure 6 st




Deformed Laminations – Deformed laminations can come from inadequate securing measures during shipment or improper handling that causes bending or pinching resulting in a permanent deformation. Permanent deformation can increase the losses of the steel. From a practical point of view, bending one lamination would not be expected to create a noticeable increase in the overall losses of a core; however, as the percentage of laminations with significant damage increases, their impact on the no-load losses will also increase.



Core Steel Grade – When the transformer is in the design stage, the core performance is typically based on a selected core steel grade. During manufacturing, if core steel grades are mixed within the same transformer core, if the properties of a given core steel grade are not within tolerance or the wrong core steel grade is used in a transformer, there may be a noticeable effect on the core performance. Increased core and other component temperatures could occur due to increases in no-load losses or slight saturation of the core. Core saturation causes increased losses in the core. Higher stray losses can also occur, since once saturation is reached in the core, flux is pushed out of the core which will create higher stray flux. Subsequently, there may also be higher losses in the tank, core clamp and other components that see higher stray flux.

FACTORY TESTS FOR DETECTING CORE PROBLEMS There are a number of factory checks or tests that are used to verify that the core design is adequate, to perform quality checks on sub-vendors, and to help identify and troubleshoot possible core problems. Some of these tests are also used in the field. 

No-Load Losses – The no-load losses are the transformer losses associated with the core, they are also referred to as core or iron losses. There are two primary components, hysteresis losses and eddy current losses, both of which occur in the core laminations when excitation voltage is impressed on the transformer. The hysteresis losses are the result of the main core flux changing direction each cycle of the power frequency, realigning the magnetic domains of the core each time. The eddy current losses are those associated with the induced eddy currents in the core laminations as a result of the main flux in the core. This test is performed by energizing the terminals of a winding with the terminals of all other winding open circuit. The losses measured at the energized winding terminals represent the energy associated with energizing the core and are the no-load losses. The no-load losses test is a standard routine test [1] so it would be performed on all new transformers. The measured no-load losses should be compared to the calculated values so that any significant difference can be investigated. It is important to identify the cause of any discrepancy to determine if a condition exists that may cause a problem later in service.



Exciting Current – Exciting current is that which flows in the excited winding terminals when all other winding terminals are open circuit. It is measured during the no-load losses test and represents all losses of the transformer when in the open circuit configuration. It is measured at rated voltage and usually several other excitation voltage levels. The exciting current is compared to calculated values to determine if there are any abnormal conditions in the core. Significant differences between the calculated and measured values should be investigated. Measurement of exciting current before and after factory dielectric tests is common to verify if any changes have occurred in the electromagnetic circuit, including the core.



Extended Over-Excitation – Since marginal defects or localized problems in a core may not be detectable by other tests, operating the core at an elevated flux density for an extended period of time has a higher likelihood of revealing core problem conditions that cause overheating and gas st


generation. They can be detected through dissolved gas analysis (DGA) tests performed before, during and after such extended tests. This non-standard test is commonly used to help trouble shoot problems and is used sparingly as a quality check. 

Lamination Resistance – Resistance between groups of laminations and to ground can be measured to detect core problems during manufacturing. Lower resistance measured across a lamination group can indicate problems such as excessive burr, improper stacking, conductive contamination and other shorted lamination conditions.



Single-Phase Exciting Current – This test is used by some manufacturers an in-process quality check. It is also performed at the factory to establish a benchmark that can be used as a reference when performing the test later in the field [2]. The test can also be used as a quality test or an investigative test when the need arises. There are some design deficiencies that can be detected by this test, as reported in reference [3]. In this case, a design deficiency caused an over-excited series transformer which was detected by the single-phase exciting current test, demonstrating that the test can be a helpful check of the transformer operating conditions. The quality and uniformity of the core construction can be determined by comparing the current and watts loss results among transformers built from the same electrical design.



Core Insulation Resistance – The core insulation is intended to isolate the core from ground, which is primarily the clamping frame in most core form power transformers. One of the critical core insulation areas in shell form transformers is at the steel T-beams. During factory testing of the transformer, the core ground insulation resistance is measured to determine if the insulation is shorted, possibly from conductive debris or a foreign object bridging the insulation to ground. It also verifies if the insulation is in acceptable condition. As a guideline, core insulation resistance of 500 MΩ is given in IEEE C57.152 [4] as a “typical” minimum acceptable value for new transformers. This recommended value may be considered for core form transformers, however some shell form transformers have lower values when measured just after final stacking, prior to tanking and the hot oil spray drying process. In some cases, lower insulation resistance values are caused by the pressboard insulation material used which is not yet dry processed and oil impregnated.

FIELD TESTS FOR DETECTING CORE PROBLEMS In service, a number of diagnostic tools are used to determine the condition of transformers and to help locate the source of possible problems. Table 1 provides a number of tests that can be effective in diagnosing a core problem. The tests’ ability to detect core problems listed in the table are marked with an “X”. There are varying degrees of severity for some problems which can impact the effectiveness of a given test to detect the problem.

Table 1 Detection Methods for Core Problems in Service Diagnostic Test DGA Capacitance Insulation Resistance FRA Exciting Current

Separated Joint(s) X

Unintentional Ground X

Core Problem Loss of Overheating1 Ground X X X

Shorted Laminations X

Core Quality2

X 3

X

X

X3

X

X3

X4

X3

X

st


Notes: 1. Overheating due to elevated temperatures in the core, for example due to poor cooling or overexcitation. 2. Core quality, for example mixing core steel grades, changes in core steel grade or incorrect core steel grades, can be detected using these tests. 3. Detectable if the problem creates a tangible change in the core reluctance encountered by the magnetic flux. This, however, is not always the case with these problems. Moreover, unless the change is very significant, the effect is often comparable with changes introduced by residual magnetism. 4. Detectable if loss of core ground significantly impacts capacitive component of exciting current. The analysis of test results will be different depending on whether the measured current is inductive or capacitive; in general, would require an understanding of the issues for proper diagnostics.

Dissolved Gas Analysis (DGA) Core problems can be detected by the gases in oil generated from overheating and electrical discharge. Core overheating can be treated as a bare metal condition since there is usually relatively little cellulose in contact with the core. If paper is involved then it is usually assumed to indicate that coils, leads or other paper insulated components are involved. Without paper, a fault would likely include components such as the tank, tap changer, core clamp, and the core. Traditionally, the ratio of carbon dioxide to carbon monoxide (CO2/CO) has been used to determine if paper is degraded due to a fault. However, it has been shown that the CO2/CO ratio is not always reliable when small amounts of paper are involved [5]. Testing for furanic compounds in the oil can also be used to identify if there is paper degradation [6] but small amounts of paper involved in a fault may not be detected by this method. When electrical arcing and extreme high temperatures are involved the furans are detected in very low quantities. The classical Duval Triangle helps differentiate high, medium and low temperature faults in mineral oil equipment. Duval [7] has also developed two triangles specifically for low temperature overheating of oil, less than about 250 ºC, which differentiate faults with and without paper involvement. The Duval Triangle 4 categorizes low temperature overheating problems in mineral oil by comparing the “low energy” gasses hydrogen (H2), methane (CH4), and ethane (C2H6). Low temperature overheating of oil, without paper involvement, is defined by a band with less than 10 % hydrogen and methane greater than about 25 %. The Duval Triangle 5, for low temperature faults in mineral oil, uses the three “hot-metal” gasses, methane, ethane, and ethylene (C2H4). Two ranges are given here for low temperature overheating of oil without paper involvement, both with ethylene less than about 10 %. One band is with ethane greater than about 50 % and the second with ethane less than about 15 %. A narrow band with ethylene less than 1 % and ethane less than 15 % is excluded, which indicates partial discharge. Hydrogen generation has been observed in low temperature overheating of cores with hot spot temperatures in the range of 120 to 160 ºC [8, 9]. The mechanism was explained as a surface area effect where hydrogen was generated on the order of 1, to over 3 ppm/day, for six identical transformers. Hydrogen gas generation was shown to be proportional to the level of over-excitation of the core and the core hot-spot temperature. The hydrogen generation trend is noted to taper off over time. Since hydrogen gas generation in oil is not expected until temperatures of about 300 ºC are reached, the effect was concluded to be due to the decomposition of oil between core steel laminations. This was explained as a catalytic effect involving the surface of the core steel, possibly due to hydrogen remaining on the surface of the laminations. Medium temperature overheating of the core can be considered to be somewhat above 160 ºC and up to about 300 ºC; as the oil overheats in a transformer, quantities of methane would be generated and then increasing amounts of ethane would be seen as the temperature increases. High temperature overheating of the core can be considered to be approximately between 300 ºC and 700 ºC where ethylene generation will start to be generated at a higher rate with lessening relative amounts of methane st


and particularly ethane. Evolution of acetylene (C2H2) is an indication of very high temperature overheating, since it is seen at temperatures above about 700 ºC. N. Nakahito and TEPCO [10] developed a method to help evaluate gas results when trying to categorize fault types; the ratio of ethylene to ethane was extrapolated from this method by the Doble Materials Laboratories and was applied to a number of cases to evaluate its acceptability. The ethylene/ethane ratio ranges given in Table 2 were used to evaluate 31 localized high temperature overheating cases, which showed that the ratio is helpful but not correct every time. Therefore, caution should be exercised when applying these ratio ranges since the results were misleading in some cases. Twenty five of the thirty one cases reviewed were diagnosed correctly with this method, showing there is favorable probability that the ethylene/ethane ratio will provide a correct indication when a high temperature overheating problem exists. In some cases, the ethylene to ethane ratio correctly showed that paper was involved and the CO2/CO ratio missed the problem. This demonstrated that in some cases, when the ethylene/ethane ratio is less than 4, paper may be involved even when the amounts of carbon oxide gases and the CO2/CO ratio does not indicate a paper overheating problem. Evaluation of the ethylene/ethane ratio is applicable to high temperature overheating so this ratio should be greater than 1 and the ratio of acetylene/ethane should be less than or equal to 0.1. The ethylene/ethane ratio is not correct every time but it has shown to be useful as an additional method to help evaluate if paper is involved.

Table 2 Ratio Ranges to Determine Paper Involvement Ratio

Paper Involved in Fault

Paper Not Involved in Fault

<4

>4

<3

>7 (Normal)

Ethylene/Ethane (C2H4/C2H6) Carbon Dioxide/ Carbon Monoxide (CO2/CO)

A shorted lamination condition can cause bare metal high temperature overheating of cores. Up until the mid-1970s, the Westinghouse shell form transformer design did not utilize solid insulation on areas of the steel T-beam, and relied on an oil gap between the T-beam and the core [11]. If shifting occurred, the uninsulated edges of the core laminations could come in contact with the metal T-beam, creating a shorted lamination condition. The gasses generated in the case of a 560 MVA, 136.8-17.2 kV, 1972, Westinghouse shell form transformer [11], demonstrate the advantage of including the ethylene/ethane ratio since the amount of carbon oxides and the CO2/CO ratio did not detect paper involvement [12]. Based on the findings of the investigative teardown and DGA results, it is likely that there was T-beam shorting of the core for a long period of time and then a second problem of shorted winding strands probably evolved in the last few years as shown in Table 3. The ethylene ratio is seen to change from greater than the 4:1 ratio indicating bare metal heating, to less than 4:1 in the later years indicating that paper was part of the overheating problem. The strands of a series winding connection were found melted open with carbon and coking in the area; paper was involved in the fault although in a small amount. It can be seen in Table 3 that the amount of carbon oxides and the CO2/CO ratio was fairly insensitive to this significant change, even though that the transformer was very close to a possible catastrophic failure.

st


Table 3 Dissolved Gas Results, High Temperature Overheating in Shell Form Transformer, [ppm] Gas Nitrogen Oxygen Carbon Dioxide Hydrogen Carbon Monoxide Methane Ethane Ethylene Acetylene Total Comb. Gas Ethylene/Ethane CO2/CO

(N2) (O2) (CO2) (H2) (CO) (CH4) (C2H6) (C2H4) (C2H2) (C2H4/C2H6) (CO2/CO)

4/27/1995 104000 732 5270 1160 60 4910 1630 4770 5 12535 2.9:1 88:1

5/17/1994 42300 2690 1340 0 8 27 32 97 0.4 164 3.0:1 168:1

4/18/1991 88200 1620 8590 10 47 235 150 578 0 1020 3.9:1 183:1

10/25/1974 80000 8900 1500 73 36 760 98 650 5 1622 6.6:1 42:1

Once it has been determined that there is a reasonable likelihood of bare metal heating, additional investigation is usually needed to determine if the core is the problem. Additional information should be reviewed, including the results of other tests, the system loading and voltage level. The system voltage level and possibly tap position will dictate the flux density in the core; higher voltage levels may increase gas generation if there is a core problem. 

Capacitance - Winding capacitance results from the overall power factor test can be used to detect changes in the winding-to-ground capacitance when prior test results, or other benchmark information, is available for comparison purposes. Under normal test conditions the core would be grounded and the CL capacitance would be a measure of the low voltage winding to adjacent grounded items, comprised primarily of the core and the high-voltage winding. Since the HV winding is guarded for this test, the current through CHL is not measured; only the current flowing to the grounded core, clamp components and other grounded items is measured. If bonding of the core to ground is not effective for some reason, then the CL capacitance will change since the core’s capacitance to ground then becomes part of the CL measured capacitance.



Core Insulation Resistance - The core ground insulation resistance test is well known and relatively easy and quick to perform in the field. As previously mentioned, this test can detect problems involving the core ground insulation. Low insulation resistance and shorts between the core and ground are known to come from shifting of the core laminations, conductive contamination, debris or foreign objects bridging the core-to-ground insulation. Since the core-toground insulation measurement is an overall measure of the entire core to ground, local changes involving a small percentage of the core-to-ground insulation may not be easily detected by this method. IEEE Standard C57.152-2013 provides values for interpreting core-to-ground resistance of transformers in service when insulated with mineral oil, see Table 4 [4]. However, it is known that different insulating materials are used to isolate the core and there can be other factors that may affect these ranges.

Table 4 Typical Core Insulation Resistance Ranges for Service Aged Transformers Insulation Resistance [MΩ] > 100 10 – 100 < 10

Indication of Insulation Condition Normal Deteriorated Investigate st




Frequency Response Analysis (FRA) - The frequency response analysis (FRA) open circuit test of transformers is sensitive to changes in the magnetic core in the lower frequency range [1316] where the magnetic flux is effectively confined to the core. The core influences the open circuit trace through the magnetizing inductance usually up to about 1 to 2 kHz. At higher frequencies, the influence of the magnetizing inductance has less effect and eventually becomes negligible. The influence of the core is not a fixed frequency range since it will vary depending on the transformer size and ratings. Changes in the core reluctance path will be detectable in the core influence range as they become significant enough to affect the response. There are other factors that influence the core in this region such as core magnetization and the effect of magnetic viscosity that should be recognized [14]. Loss of core grounding can be detected in the higher frequencies, for example 50 kHz and higher, where changes to the CL capacitance would be reflected [16].



Exciting Current – The single phase exciting current test is performed with the secondary windings open circuit so the magnetic flux is primarily confined to the core, a condition similar to the FRA open circuit test. In service, this test can be a useful tool to detect core problems that are related to shorted laminations and other problems that significantly affect the reluctance of the flux in the core; such as a partially shifted or open core joint [17].

CORE SERVICE PROBLEMS Core problems encountered in service include core ground problems, shorted laminations and overheating. There are differences between the factory and service environment, although some of the testing and diagnosis are very similar. One advantage, once the transformer is in service, is that there should be benchmark data available that can make interpretation of the field test data more conclusive.

Core Ground Most modern power transformers are designed with externally accessible means for electrically bonding the core to ground at a single point while in service. If more than one core ground is present while in operation, circulating currents may develop causing overheating and gassing. A number of papers have been written addressing unintentional core grounds, means of detecting and resolving them [18-20]; therefore this topic will be addressed briefly. Unintentional core grounds may not be severe in nature, at least to start with, but should be addressed. A circulating current path through core laminations can cause high temperatures, which can further damage the core and insulation, creating a condition that deteriorates over time. Gas generated due to high temperatures may also mask other possible problem conditions in the transformer. An example of gas generation due to an unintentional core ground problem found on a 53 MVA, 345 kV, GSU transformer is given in Table 5.

st


Table 5 Gas Generation of Unintentional Core Ground Gas Oxygen Nitrogen Hydrogen Methane Ethane Ethylene Acetylene Carbon Monoxide Carbon Dioxide CO2/CO ratio Ethylene/Ethane ratio Gassing rate, TCG, ppm/day

12-2-1992 25000 70000 63 96 31 240 2 630 6200 9.8:1 7.7:1 0.3

9-23-91 22000 79000 31 21 12 100 1 650 6000 9.2:1 8.3:1 -

Loss of Core Ground While a single point core ground is necessary in most designs, larger cores may have multiple sections that are isolated from one another by cooling ducts to help maintain acceptable operating temperatures. Depending on the design, these ducts can divide the core into electrically isolated sections, where additional means are needed in to ensure grounding of all sections of the core. Internal core jumpers are often used to bond the sections of the core together and then a single lead can be routed to the core ground bushing. Means must be provided in the design to ensure adequate pressure is maintained on the core section jumpers and the core ground lead electrode to keep all parts of the core properly grounded throughout the service life of the transformer. If the core bonding to ground is ineffective for any reason, the core can be charged through the coil-tocore capacitance. The core is capacitively coupled to the windings, primarily to the inner most coil; usually the LV winding in a two winding transformer. With no core ground, the core can charge up and then discharge when the potential difference is great enough to bridge the core insulation to ground. If the core is not bonded to ground, there can be a noticeable change in capacitance between the inner coil and ground, which may be detected when performing the overall power factor and capacitance test on the windings. Figure 7 shows an example where the continuity of the core ground was intermittent, since the CL capacitance changes several times between two capacitance values. The core was determined to be ungrounded for the tests performed in 1993, 1994 and 2005. The trend to note in this case is that the capacitance returns to its original value which would not be expected in the case of winding deformation. This transformer was moved at the site, which could help explain, at least in part, the cause for losing the core bonding to ground. A loose connection was suspected; one explanation is that the core laminations were not compressed tight at the location where the core ground electrode was placed in the core, allowing the ground electrode to lose contact with the core.

st


Capacitance Pattern of Intermittent Core Ground Figure 7 DGA results showed that overheating of bare metal was occurring in the transformer main tank, a sample from 2004 is given in Table 6. The ethylene to ethane ratio indicates that the heating was without paper insulation involvement. The carbon oxide ratio is at the limit indicating no paper involvement.

Table 6 Gas Generation of Unintentional Core Ground Gas Oxygen Nitrogen Hydrogen Methane Ethane Ethylene Acetylene Carbon Monoxide Carbon Dioxide CO2/CO ratio Ethylene/Ethane ratio

7/25/2004 845 16,498 75 159 30 174 1 613 4,027 6.6:1 5.8:1

Core Overheating In service, core overheating can be due to a number of problems, including poor cooling, over-excitation and shorted laminations. Core defects such as shorted laminations can occur due to a number of causes, including damaged or deteriorated lamination insulation, and debris or conductive foreign objects that come in contact with the un-insulated edges of the core. If the cause of the shorted laminations also bridges the core-to-ground insulation, it may be possible to detect these defects by performing a core-toground insulation resistance test. If a defect only bridges core laminations, and does not involve ground, then other tests will likely be needed to identify the problem.

st




Shorted Laminations - The following example details a shorted lamination condition that was identified following initial energization of the transformer. A new 3-phase, 350 MVA, 230 kV class, generator-step-up (GSU) transformer showed an unusual gassing pattern within hours after being placed into service. Data from the on-line gas monitor showed a hot metal gassing pattern, see Figure 8. The on-line values were confirmed with laboratory tested samples.

On-Line Monitor Gassing Pattern Following Initial Energization Figure 8 The transformer was returned to the factory where investigative testing was performed which included an extended over-excitation test. The over excitation test was performed at 110 % for the first 24 hours and then at 115 % for an additional 4 hours. The general gassing pattern was similar during the investigative test as shown by comparison of individual gasses in Figures 9 and 10. It can be seen that the ethylene and acetylene follow similar trends, although the generation rates for the factory over-excitation test results are less consistent. The extensive factory investigative testing showed that the gas generation was dependent on excitation voltage.

Ethylene Gas Pattern Figure 9

st


Acetylene 3.0 2.5 2.0 Conc, ppm 1.5

28hr OE C2H2

1.0

Sev Field C2H2

0.5 0.0 0

10

20

30

Duration, hrs

Acetylene Gas Pattern Figure 10 A controlled investigative teardown was performed to find the root cause of the problem. After all of the findings of the teardown were reviewed, it was concluded that a relatively small burn mark, approximately 1 inch long, found on the core lamination edges was the cause of the gassing, see Figure 11.

Core Lamination Edge Damage Figure 11 During the teardown, the core steel was removed in small groups and then restacked on a pallet for temporary storage. The photographs in Figure 11 were taken after the core steel was removed so the order of the laminations and the damage mark are not the same as when the laminations were installed in the transformer. Small pit marks and carbon were observed in the damaged area, see right photo of Figure 11. The laminations were effectively welded together at the burn mark since some force was required to separate them for inspection. The mark was located on the bottom side of the top yoke. This location made it less likely that debris or a foreign object came to rest on the lamination edges and caused the shorted condition. It was concluded that lamination edges were shorted together due to damage of the lamination insulation at the edges that allowed a path for circulating currents which caused the gassing. Although this core problem was found during start-up, it is possible that this damage area was present during manufacturing. This type of core problem is not always detectable during manufacturing; if it is not identified and fixed during construction at the factory it would normally only be revealed once full st


excitation is impressed on the transformer for a period of time. This case reveals a weakness in the factory testing standards since the IEEE standard test regimen was followed with respect to the core. As required by the standards, transformers are at full excitation during the no-load losses test that is typically conducted for a period long enough to record the required parameters, which is usually several minutes. 

Local Overheating - Localized overheating was found on a single phase GSU, shell form transformer, rated 500 kV, 130 MVA, with FOA cooling. It was likely due to shorted laminations, see example in Figure 12. The left photo shows the overheated area on the outside edge of this shell form core. The right photo shows the depth that the overheating damage penetrated into the core. The transformer was nearly 50 years old at the time of decommissioning.

Core Overheating at Edge Figure 12 

Blocked Oil Flow - Figure 13 shows the result of a suspected partially blocked oil duct on the same shell-form transformer core that was discussed in the previous example. Based on the teardown investigation, it was concluded that a wood shim driven between the core and tank, at the time of manufacturing, extended past the core and effectively blocked oil circulation in this oil duct. The insulation was dark at the blocked duct, see left photo in Figure 13. The core lamination at this duct was found bent near the edge, which likely occurred during installation of the wood shim, see right photo in Figure 13. The lamination at this duct showed signs of significant overheating since the surface was black and was rough with a granular texture.

Core Overheating at Cooling Duct Figure 13 st




Joint Overheating - Overheating at a core joint is shown in Figure 14. This was found during the decommissioning teardown of a three-phase, GSU, shell-form transformer rated 500 kV, 390 MVA, with FOA cooling. A significant winding fault condition likely masked any gas generated from the core in this case. Over-excitation of the core was the suspect cause of this overheating condition.

Core Joint Overheating Figure 14 Infrared thermography is a helpful tool that can be used to verify core corner joint heating in shell type transformers. The shell form transformer core is usually located near the tank, typically less than one inch, and experience shows that the corner joints are more susceptible to have larger gaps. Figure 15 shows a thermal scan from a 66 MVA, 140.9 – 34.5 kV, shell type transformer, with the highest temperature found at the left corner of the core.

Thermal Scan Showing a Hot Spot at the Left Corner of the Core Figure 15 Separation of Core Joint(s) The magnetic core of the power transformer provides a low reluctance path to the magnetic flux to help maximize the efficiency of linking the primary and secondary coils. The many thin laminations of the core st


are commonly held in place with a number of insulating blocks placed between the core and clamping frame to prevent the laminations from moving during lifting, transportation, installation and service. If the joint of a core becomes separated to the point that the magnetic flux is not able to easily bridge between adjacent core laminations, then the reluctance will increase since the localized saturation area is increased. It is not that common for core joints to open up in service and may be more likely to occur during transportation or when moving if significant forces are encountered. After shipment of a three phase, 34.5 kV, 90 MVA, rectifier transformer to site, questions were raised due to several moderately high impacts, up to 4 G, recorded during the railroad segment of transportation. A limited internal inspection at the site revealed that several core blocks were shifted out of place and one was completely displaced and lying on the bottom of the tank. It was determined that the transformer would have to be returned to the factory in order to fully inspect its internal assembly and make the necessary repairs. Once the transformer arrived back at the factory, electrical acceptance tests were repeated to continue investigating the condition of the transformer. The testing revealed that the no-load losses had increased from 73.1 kW, measured during the initial factory test, to 76.2 kW; an increase of more than 4 %, see Table 7. Exciting current during the no-load loss measurement increased from approximately 1.6 to 2.2 A, an increase of nearly 40 %. The manufacturer considered an increase in no-load losses of 1 % or less to be acceptable. Since the increase in the measured no-load losses were greater their 1 % limit, the results were considered to indicate a problem more extensive than the loose and displaced core blocks.

Table 7 No-Load Losses and Exciting Current at 100 % Rated Voltage Factory Acceptance Test

Factory Investigative Test

No-Load Losses, [kW]

73.1

76.2

Exciting Current [A] at 100 % voltage

1.6

2.2

FRA test results showed a shift in the core region, at about 5 kHz and lower in this case, see impedance plot in Figure 16. The relatively small change is comparable to the effect of residual magnetism. If similar results are encountered and a problem is suspected with the core, demagnetizing the core and repeating the test may help rule out residual magnetization.

FRA Shift in Core Response Range Figure 16 st


The active assemblies were removed from the tank, revealing that one of the cores had sustained damage. A large gap was noticeable at a top core joint indicating that the joint was separated, see Figure 17. Other joints on the core had partially separated. Several bottom core yoke blocks were completely displaced, allowing the laminations to bend. This created several large gaps between laminations, Figure 18, further indicating that the core was not suitable for service.

Gap

Separated Core Joint at Top Yoke Figure 17

Missing Core Block

Missing Core Block

Gaps Due to Bend in Laminations

Core Laminations Bending and Gaps Observed Figure 18 CONCLUSION There are a number of diagnostic tools available to help detect and determine the nature of core defects. Dissolved gas analysis can detect a wide range of core issues, where differentiation of the core and other possible sources becomes the challenge once a gassing issue is identified. A number of electrical tests discussed here, including overall capacitance, core ground insulation resistance, FRA and exciting current, can be used to help narrow down the possible type of core problem that may exist on a given transformer. It is usually more effective to apply as many relevant tests as possible to provide detailed st


data so that decisions on how to resolve possible problems can be based on the best possible information.

REFERENCES [1] IEEE Standard C57.12.90, IEEE Standard Test Code for Liquid-Immersed Distribution, Power, and Regulating Transformers, IEEE, 3 Park Avenue, New York, NY 10016, USA. [2] Lachman, M. F., “Exciting Current and Losses”, Tutorial Presented at the Eightieth Annual International Conference of Doble Clients, Boston, MA, 2013. [3] Payne, P. A., “Why Specify Single Phase Low Voltage Excitation Test for Transformers as a Production Test?”, Proceedings of the Sixty-Ninth Annual International Conference of Doble Clients, Boston, MA, 2002. [4] IEEE Standard C57.152 - 2013, IEEE Guide for Diagnostic Field Testing of Fluid-Filled Power Transformers, Regulators, and Reactors, IEEE, 3 Park Avenue, New York, NY 10016, USA. [5] Griffin, P. J., “Discussion of the Bert R. Hughes and Harold Moore Paper, ‘Analysis of Gasses in a Large GSU and Special Operating Guidelines for a Family of Large Westinghouse Transformers, Manufactured Without Insulated T-Beams Generating High Levels of Combustible Gasses’”, Proceedings of the Sixty-Third International Conference of Doble Clients, Boston, MA, 1996. [6] Lewand, L., “Practical Experience Gained from Furanic Compound Analysis”, Proceedings of the Seventy-Third Annual International Doble Client Conference, Boston, MA, 2006. [7] Duval, M., “The Duval Triangles for LTCs, Alternative Fluids, and Other Applications”, Proceedings of the Seventy-Sixth Annual International Doble Client Conference, Boston, MA, 2009. [8] Oommen, T. V., Girgis, R. S., Ronnau, R. A., “Hydrogen Generation from some Oil-Immersed Cores of Large Power Transformers”, Proceedings of the Sixty-Fifth Annual International Conference of Doble Clients, Boston, MA, 1998. [9] Girgis, R. S., teNyenhuis, E. G., “Hydrogen Generation Due to Moderately Heated Transformer Cores”, Proceedings of the Seventy-Third Annual International Doble Conference, Boston, MA, 2006. [10] Nakahito, K., “Recent Activities on Maintenance Management for Power Transformers”, Proceedings of the Seventy-Fourth Annual International Doble Client Conference, Boston, MA, 2007. [11] Hughes, B. R., Moore, H., “Analysis of Gases in a Large GSU and Special Operating Guidelines for a Family of Large Westinghouse Transformers, Manufactured without Insulated T-Beams Generating High Levels of Combustible Gasses”, Proceedings of the Sixty-Third International Conference of Doble Clients, Boston, MA, 1996. [12] Lewand, L., R., Griffin, P., J., “Case Studies Involving Insulating Liquids and Materials from the Doble Materials Laboratories”, Proceedings of the Seventy-Seventh Annual International Doble Client Conference, Boston, MA, 2010. [13] Lachman, M. F., Fomichev, V., Rashkovski, V., Shaikh, A. M., “Frequency Response Analysis of Transformers: Visualizing Physics Behind the Trace”, Proceedings of the Seventy-Eighth Annual International Conference of Doble Clients, Boston, MA, 2011. [14] Lachman, M. F., Fomichev, V., Rashkovski, V., Shaikh, A. M., “Frequency Response Analysis of Transformers and Influence of Magnetic Viscosity”, Proceedings of the Seventy-Seventh Annual International Conference of Doble Clients, Boston, MA, 2010. st


[15] Abeywickrama, N., Serdyuk, Y. V., Gubanski, S. M., “Effect of Core Magnetization on Frequency Response Analysis (FRA) of Power Transformers”, IEEE Transactions on Power Delivery, Volume 23, Number 3, July 2008. [16] IEEE Standard C57.149, IEEE Guide for the Application and Interpretation of Frequency Response Analysis for Oil-Immersed Transformers, IEEE, 3 Park Avenue, New York, NY 10016, USA. [17] Freund, D. A., “Transformer Damage Detected by Excitation Current Tests”, Proceedings of the Forty-Third Annual International Conference of Doble Clients, Boston, MA 1976. [18] Gamache, M., Pong, L., “Unintentional Transformer Core Ground – Diagnostic and Mitigation”, Proceedings of the Seventy-Eighth Annual International Conference of Doble Clients, Boston, MA, 2011. [19] Bates, D. E., Brunson, P. W., “Detection, Analysis, and Rehabilitation of Unintentional Core Grounds in Large Power Transformers”, Proceedings of the Fifty-Eighth Annual International Conference of Doble Clients, Boston, MA, 1991. [20] Phillips, R. H, “Transformer Core Grounds”, Proceedings of the Fortieth Annual International Conference of Doble Clients, Boston, MA, 1973.

BIOGRAPHY Bill Griesacker is a member of Doble Global Power Services, employed as a transformer engineer working on projects that include factory inspections, condition assessment, design reviews, failure analysis and other general consulting projects. He previously worked for Pennsylvania Transformer Technology Inc., where he held various positions including Engineering Manager. His work included high voltage insulation design, transient voltage modeling of power transformer windings and various LTC and DETC switch development projects. Prior to this, he was employed by the Westinghouse Electric Company, working on synchronous generator projects as a member of the Generator Engineering Department. Bill started his career with Cooper Power Systems in large power transformers and later worked in the Kyle Switchgear, Vacuum Interrupter Department. He has earned a M.S. degree in electric power engineering from the Rensselaer Polytechnic Institute and a B.S. degree in electrical engineering from Gannon University. Bill is an active member of the IEEE, PES Transformers Committee where he holds positions in several working groups and subcommittees. Juan Luis Thierry is a Senior Transformer Consulting Engineer for Doble Consulting and Testing Services, concentrating on transformer design review, factory inspections, factory acceptance tests, forensic analysis and condition assessment; where he applies his more than 30 years of power transformer experience in QA, manufacturing and electrical design for both transformer types, shell-form and core-form. Prior joining Doble Engineering in September 2009, Juan Luis held positions as Engineering Transformer Platform Leader, Engineering Manager Transformer Services and Shell Design Engineer Team Leader from 2004 to 2009 in GE Bradenton Transformer Service Center. He was also Engineering Manager, Shell Section Engineering Manager, Fellow Design Engineer, Principal Design Engineer and Senior Design Engineer from 1998 to 2004 in Ohio Transformer Bradenton Remanufacturing Center in Florida, where he worked in the electrical design of the remanufacture of shell type three-phase transformers up to 1000 MVA and single-phase transformers up to 600 MVA, in voltage class up to 525 kV and 1675 kV BIL. Since 2004, he initiated the manufacture of new shell transformers back to USA in the Bradenton facility, designing both the mechanical and electrical designs for three-phase transformers up to 800 MVA and single-phase transformers up to 200 MVA, in voltage class up to 525 kV. He designed more than 85 shell type transformers during his tenure in Bradenton, Florida.

st


Before coming to USA, Juan Luis work in IEM-Westinghouse; a facility outside Mexico City, from 1982 to 1998, where he held positions as Chief Design Engineer, Design Engineer, Manufacturing Chief and QA Auditor. He worked in the electrical design for new and remanufactured transformers in both types of core construction, core-form and shell-form. Juan Luis received his Bachelor of Science in Electrical and Mechanical Engineering from the National Autonomous University of Mexico (UNAM) in 1981, where he graduated with honors and received the “Gabino Barreda Medal” granted by the UNAM in 1982 by obtaining the best GPA of the class from 1978 to 1981. Juan Luis is a member of the IEEE Power Energy Society Power Transformer Subcommittee and member of the IEEE Standards Association. He is a member of the U.S. National Council of the International Council on Large Electric Systems (CIGRE). He is also a member of the American Society of Mechanical Engineers.

st


Transformer_Magnetic_Core_Issues.pdf

Recommend Documents